By its nature, the internet is a system of interconnected networks. These interconnections rely heavily on the hyperlink acting as a bridge between independent sources of information. In large part the value of the internet is derived from the quality of the information and the optimal linking of useful and complimentary information sources.
The history of the worldwide web has been shaped in large part by the companies and organizations that can most effectively leverage these information resources. Common examples of this strategy include search engines that seek to organize and order indexes by relevancy to search terms. Primarily, search engines rely on machine learning to create quality scores that relate potential search queries to sources available on the web. This system has proven extremely effective under the context of a user entering a particular query and being provided with a valuable stable of organized and indexed results. On the other hand, while a search engine may excel at providing a list of sources with high relevancy for a particular search term, this index is assembled in a vacuum relative to the content it may be used in. The top result for a particular search term may not be the best source to place in a particular piece of content.
Web content is a human creation. Producing a blog article, developing a website or posting a link on social media are subjective, personal acts. Generally, these acts of creation seek to appeal to and share with a particular audience. In many cases, this act of sharing involves a specific source or is buoyed by a series of complimentary sources in the form of a hyperlink. Appealing to a particular audience is a primary goal of content creation, but current methods for researching source information inherently rely on a definition of relevancy based on the macro-audience of all internet users. In many cases, content creators provide a more informed selector of sources for their particular work, relying first on search engines and then on their digression to choose a final result. In all cases, the resources that are included as links in a particular piece of content are incorporated as educated guesses as to the particular sources that will most effectively appeal to their audience.
Social media is driven by human interaction. Many services rely on users voting on content in order to determine its rank and likelihood of being viewed by the general audience. A popular social media platform marketed under the trademarks REDDIT and ASK REDDIT is an example of this standard in social media sites. While sites like the REDDIT website do a particularly good job of surfacing the most interesting content submitted to the site, this model of submission and voting is flawed. The vast majority of users on sites of this nature avoid the “active interaction” of voting for content and merely consume the sources that have been influenced by a small percentage of the user base. Ultimately the content being served and selected is not derived from the lurking users who account for the majority of the audience.
Another example of user input affecting content is A/B testing, marketed under the trademark GOOGLE. This feature allows a web developer to set up competing landing pages on the same website. These pages are routed and served at random to different web users by analytics. As a segment of the users interact with page A and a segment interacts with page B, their behavior is measured against a common conversion standard. Once the testing period has served the pages to a large enough sample size of users to determine the statistically most effective page relative to the conversion goal, the service can permanently serve this page to all future users.
Another service that does not involve user input or interaction, but that returns a single result is that marketed under the trademarks I'M FEELING LUCKY and GOOGLE. This search functionality allows a user to enter a search query, click the “I'm Feeling Lucky” button and be randomly taken to a single relevant result in the search set for that term. Although this service may, in some cases, simplify basic searching, it represents no efforts toward suitability for a specific audience or application, but rather is an offshoot of the same relevancy functionality of a basic search engine. The “I'm Feeling Lucky” feature is truly one dimensional and involves no deeper level measurements or functionality.
As discussed above, current examples of social voting rely on multiple users posting resources to an aggregator site where the audience votes for those resources only in the context of the aggregator site's culture/environment. This process fails to incorporate the influence of the content and the specific audience of the content in which those links will eventually be reposted. A particular resource might make a great top result in the context of an active social voting site, but that same resource may be a poor fit for the audience of a niche blog or a person's personal social media page. The social voting sites also fail to capture the influence of the passive, non-voting consumers of the resource that, in fact, make up a significant majority of the audience. Moreover, while search engine results use multiple factors to determine relevancy and rankings for the results they serve in the context of a user entering a query and interacting with resources presented for that query, this process fails to tailor relevancy for a resource within the context of a particular piece of content being viewed by a particular audience. For example, a current news story, “Canyonlands National Parks Ban Drones,” may be the first search engine results for the news search query “drones,” but the result “Legislature mulls curbs on use of aerial drones by paparazzi” might be a better fit for a news blog whose audience enjoys entertainment stories. As content creators are generally responsible for selecting a resource from a search results set to best fit their content, the results of typical social media sites and search engines may require significant additional review before reliable resource selection.
Therefore there is a need for a service that provides a single “best fit” reference for a search query where the determination for the best fit reference is driven by user input from a specific subset of internet users.